►
From YouTube: K8s SIG Docs Meeting 20200211
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
It
appears
that
our
agenda
has
been
hacked
and
the
previous
one
destroyed
root
causes
are
still
totally
unknown,
but
the
document
has
been
not
only
its
contents
removed,
but
the
document
itself
destroyed,
so
we've
cobbled
together
a
temporary
agenda.
The
link
is
in
the
zoom
chat,
so
we
have
a
new
visitor,
Adam
B
Kaplan,
who
thankfully,
was
able
to
had
already
loaded
the
talk
and
was
able
to
capture
the
buffer
and
is
going
to
send
it
to
me
to
repopulate
this
new
version
of
the
doc.
A
A
A
A
As
always
make
sure
you
know
if
you're
an
approver
for
cig
docks,
make
sure
you
know
what
your
scheduled
PR
English.
It
is
alright,
let's
go
on
to
our
agenda,
so
we
have
a
piece
of
process
first,
and
that
is
a
chair
transition
in
cig
docks,
so
jennifer
Rondo
has
been
our
co-chair
for
golly
a
while.
Now,
when
did
you
first
start
Jennifer
I.
C
Actually
sometime
I
think
in
2018,
you
all
must
have
invited
me
to
become
chair,
and
last
year,
I
started
a
new
role,
which
meant
that
my
ability
to
fulfill
my
chair
responsibilities
diminished
significantly
and
it
became
critic
pretty
clear
by
the
end
of
the
year.
That
I
just
could
not
continue
to
serve
the
community
well
in
this
role
and
that
we
also
had
some
really
stellar
people
stepping
up
to
the
role.
So
it's
a
good
thing.
Well,
it's
a
it's
a
sad
thing
for
me,
because
this
is
an
amazing
community.
C
A
C
Just
one
thing
to
add
to
this
shameless
plug
one
of
the
things
that
is
particularly
pressing
right
now.
That
means
I
really
need
to
step
away
from
the
kubernetes
community
for
a
bit
is
I
am
increasingly
involved
with
helping
run
their
rights
at
our
conferences
and
I'm
in
the
process
of
looking
for
a
venue
for
an
East
Coast
Conference
this
year,
so
no
promises.
C
Yet
you
know
we
we
hope
to
do
it,
but
it's
still
a
work
in
progress,
and
we
can't
say
for
sure
that
it's
going
to
happen,
but
since
y'all
are
Doc's
at
least
Doc's
adjacent,
if
not
Doc's
immersed
and
a
lot
of
you
are
ducks
immersed,
just
reminding
you,
the
right
to
ducks
community
is
out
there.
There
have
been
some
really
fruitful
collaborations
between
the
kubernetes,
Doc's
and
rights
of
Doc's
communities
over
the
years
and
I
fully
intend
to
continue
to
push
on
those.
A
D
Hi
everyone
I'm
Caitlin
Zach,
said
I've
been
shadowing
for
co-chair
for
a
while
now,
but
I've
been
involved
in
save
dogs
for
a
long
time
before
that,
helping
with
a
lot
of
the
website
stuff,
as
well
as
leading
the
blog
sub-project
and
that's
kind
of
what
I've
been
involved
in.
A
Thank
you.
So
the
piece
of
governance
that
we
have
to
do
today
is
to
formally
accept
Jennifer's
resignation
and
then
formally
vote
in
Caitlyn
as
co-chair,
and
unfortunately
this
is
a
bit
of
theater,
since
the
only
people
voting
are
Jim
and
myself
and
Jennifer
for
the
first
one.
So
let
the
Kabuki
unfold,
doo,
doo,
doo,
emeritus,.
A
A
F
A
A
F
A
All
right
docks,
so
I
had
put
on
the
agenda
the
question
of
how
to
discuss
deepening
the
pool
of
good
first
issues.
We
have
got
a
lot
going
on
today,
so
I'm
gonna
bump
that
way
back.
We
will
get
to
that
if
we
have
time
and
I'm
gonna
put
Adams
agenda
item
before
mine
all
right
so
released.
1.18
sadita
said
that
she
would
probably
not
be
able
to
make
this
meeting,
but
she
did.
Let
us
know
that
things
are
going
pretty
well,
as
expected.
I
just
noticed,
there's
a
hideous
string
instead
of
my
actual
name.
A
There
I
fixed
it
so
two
things
that
I
noted
from
the
release,
one
that
18
process
beneath
has
done
these
unfortunately
rebased
release
that
1.17
on
to
master,
so
that
is
that
went
well.
That
would
find
so
at
this
point
I'm,
confident
that
it
will
continue
to
go
well
and
he
also
opened
the
first
merge
PR
to
merge
master
into
dev
118
and
that
also
went
really
well.
174
commits
no
conflicts,
so
that
process
is
looks
like
it's
well
in
hand.
Also,
so
1.18
I
haven't
heard
any
concern.
A
A
Basically,
the
long
story
short
is
that
there
are
a
couple
of
different
options
in
Hugo
for
declaring
the
URL
path
for
a
particular
file.
They
all
have
different
effects.
Some
of
them
are
not
very
friendly
and
it's
not
always
immediately
obvious.
What
is
the
the
correct
attribute
to
declare
in
the
front
matter?
A
So
there
are
the
three
attributes
that
are
available
are
URL,
slug
and
alias
or
aliases.
Technically
we
discovered
the
hard
way
that
URL
sets
an
absolute
path
for
a
page
without
creating
a
redirect
if
it's
an
existing
page,
so
the
takeaway
from
that
is
don't
use.
Url
use
aliases
unless
you
are
absolutely
sure
what
you're
doing
so.
A
A
So
just
be
very
careful
about
how
you're
declaring
how
you're
declaring
a
path
for
a
file
you
consider
URL
the
dangerzone
use
aliases
wherever
possible
aliases
will
it
does
exactly
that?
It
creates
an
alias,
and
it
also
sets
up
a
redirect
that
preserves
the
search
ability
and
the
discovery
of
the
discoverability
of
a
particular
file.
A
So
it
was
basically
we
have
a
you
break
it.
You
fix
it
policy
in
cig,
docs
Jim
broke,
it's
Jim's
responsibility
to
fix
it,
but
also
I
as
an
approver
I.
Let
that
change
go
like
I
know
how
these
attributes
work
and
it
still
slipped
past
me.
So
it's
my
responsibility
to
fix
as
well.
That's
the
one
of
the
things
that
we
ask
as
an
approver.
If
you
approve
a
breaking
change,
help
fix
it.
So
that's
what
I
was
doing
racing
the
clock
against
my
ambient
at
12:30
a.m.
A
to
try
and
get
a
unconscious,
so
the
takeaway
from
that
is
make
sure
that
you're
testing
your
changes,
if
you
test
like
test
locally
test
in
the
metal,
if
a
staging
build,
make
sure
that,
if
you're,
if
you're
playing
around
with
site,
architects
really
with
anything,
make
sure
that
you're
actually
testing
out
the
functionality,
so
it's
really
am
harrassing.
When
your
co-chairs.
A
G
So
I've
got
a
couple
of
questions
about
this,
because
this
kind
of
dovetails
into
the
mire
that
I'm
in
with
contributions-
and
you
answered
one,
which
is-
we
recommend
you
using
alias
unless
you
know
what
you're
doing
the
second
question.
This
is
more
just
of
curiosity,
which
is
have
we
considered
dropping
a
blank
template
file
into
the
kubernetes
directory
somewhere,
so
that
people
know
roughly
a
what
frontmatter
to
use
if
they're
doing
something.
You
know
that's
kind
of
question
number
one.
G
G
H
H
F
Right,
first
of
all,
I
do
want
to
take
you
know:
ownership
I
think
it's
good
to
have
these
discussions
in
a
larger
group
and
say
here's
what
happened.
Here's!
What
went
wrong
you
know,
and
so
this
was
initially
a
change
that
I
made
and
I
did
test
locally.
This
is
something
that
Zack
and
I
were
talking
about.
Giving
me
the
same
situation.
A
thousand
times
would
make
the
same
a
or
a
thousand
times
what
I
did.
F
What
I
did
wrong
here
was
I
tested
the
change
for
what
I
was
peeing,
so
I
wanted
to
create
the
alias
for
our
/r
bag
and
I
tested.
I
said
hey.
This
works
where
I
made
the
big
mistake
was
not
testing
previous
functionality
and
that's
my
takeaway
in
my
lesson
here
is
not
only
testing
for
what
your
PR
is
changing
but
test
for
what
was
previously
as
a
standard
for
your
PR.
So
in
my
case,
this
alias
wallet
does
work.
F
How
does
it
affect
the
existing
content
where
it
lives,
which
I
would
have
never
ever
tracked
in
a
million
years?
But
now
I
will
do
that
and
check
retro
actively
so
to
further
elaborate
on
your
questions
for
lest
I
test
locally
using
Hugo.
So
in
the
website
repository
you
can
type
in
Hugo
serve.
We
have
some
documents
around
that
I
believe
the
current
model
on
the
on
the
the
case.
Slash
website
is
using
a
docker
container.
It's
a
containerized
Hugo,
basically
to
host
your
webpages
I
think
there
could
be
an
improvement.
F
What
we
mean
when
we
say
test
locally
I
think
everyone
has
their
own
way.
Net
Liffe
ID
does
have
freeze
free
services
where
you
can
test
your
changes
directly
in
lfi
running
your
own
website,
so
I
don't
think
that
there's
necessarily
a
single
way,
but
there
is
guidance
out
there
and,
like
I,
said
personally,
I
use
Hugo
serve
locally
to
test
that
to.
A
Answer
your
first
question
and
thank
you
Jim
to
answer
your
first
question
Celeste.
Is
there
a
template?
Yes,
but
I
think
they're
looking
at
the
directory
where
they
reside,
they
are
woefully
out
of
date,
so
we
could
definitely
update
the
templates
for
the
record
templates
reside
in
website
/
layout,
/
partials,
/
templates.
D
D
G
A
G
G
A
B
Muted,
alright
yeah.
So
what
started
as
just
a
comment
between
me
and
the
team
that
owns
the
Oakland
ships,
CIA
infrastructure,
on
what
quality
of
service
meant
and
how
that
actually
affects
pod
behavior
turned
into
something
a
little
bit
larger
and
I'd
like
to
actually
share
my
screen.
If
that's
okay
with
you.
B
Alright,
so
can
you
see
it
so
as
I
started?
Reviewing
the
documentation
around
QoS
and
pot
eviction
and
scheduling.
I
grew
a
little
frustrated
in
terms
of
how
things
are
currently
laid
out
today,
particularly
in
the
concepts
section.
So
what
I'm
proposing-
and
this
might
be
something
that
I
guess
we
can
discuss
how
I
could
move
this
forward.
B
So
I
spent
a
day
splitting
it
into
very
I
think
it's.
The
list
is
now
10
10
items
long
in
terms
of
things
to
do
some.
Some
of
it
is
moving
around
existing
documentation,
but
others
is
proposing
new
documentation.
So
I
don't
know.
If
folks
would
like
me
to
stop
here
or,
and
maybe
explain
a
little
bit
more,
where
I'm
coming
from
or
keep
moving
forward,
I.
B
The
focus
is
going
to
be
helping
people
who
are
new
or
experienced
to
understand
how
workloads
are
assigned
to
pods
how
pods
can
be
guided
to
specific
nodes,
how
they
can
avoid
and
then
go
into
what
happens
when
your
cluster
is
under
load,
which
is
when
the
eviction
piece
comes
in.
So
some
bits
like
assigning
two
pods
or
nodes
that
are
from
nodes,
is
currently
under
the
configuration
section
which,
if
you
look
at
that
today,
it's
almost
like
the
junk
drawer
in
your
kitchen
or
your
office.
There's
just
a
bunch
of
miscellaneous
stuff.
B
There
I
was
proposing
moving
that
from
configuration
to
scheduling
an
eviction
moving
the
tapes
and
Toleration
x'
article
to
scheduling
an
eviction
moving
pod
priority
and
prevention
from
configuration
to
scheduling
an
addiction
with
some
modifications
proposing
new
articles
around
the
quality
of
service
concepts,
which
I
believe
it's
currently
in.
There's.
There's
some
items
there
under
tasks
as
well
as
creating
a
new
article
under
fiction
and
then
moving
a
few
other
bits
that
are
currently
under
conviction,
configuration
or
the
scheduler,
moving
them
into
scheduling
and
eviction,
but
retitling
them
as
advanced
topics.
B
Some
of
them,
particularly
around
scheduler
performance
tuning
resource
bin,
packing
of
the
scheduling
framework,
are
also
features,
and
these
are
things
that
are
not
enabled
by
default,
or
are
things
that
you
know
you're
only
gonna
really
look
at
this
if
you're
really
digging
into
the
weeds
or
you
want
to
be
on
the
bleeding
edge.
So
the
thought
process
is
that
we
kind
of
move
from
the
scheduler
introduction
working
through
the
logic
of
how
the
your
pot
is
placed
with
the
assigning
pots.
B
E
So
I
really
like
this,
you
know
scheduling
an
eviction.
Is
you
know
those
those
are
very
obviously
complex
topics
and
the
fact
that
you
know
you
have
things
that
are
in
configuration,
but
moving
them
all
to
that
umbrella.
Topic
of
scheduling
an
eviction
I
think
would
really
help
folks
find
everything.
That's
related
to
those
topics
and
Adam
I.
A
D
B
A
A
Then
open
a
series
of
PRS
just
start
opening
PRS
that
actually
implement
the
bits
here
and
we
have
really
good
reviewers
and
folks
who
I
think
would
love
to
see
this
content
happen.
So
yeah
I
get.
We
can
give
you
plenty
of
feedback
and
review
to
make
sure
that
the
content
makes
sense
and
that
it
plugs
in
well
into
the
website,
and
we
can
even
help
you
avoid
the
pitfalls
of
frontmatter,
attribute
declarations
so
but
yeah.
A
It
sounds
like
this
might
be
like
a
like
an
umbrella
issue
with
a
whole
bunch
of
PRS
underneath
it,
and
one
thing
to
consider
is
whether
you
want
to
do
those
whether
you
want
to
do
an
iterative
process
for
you
like
merge
PRS
as
you
go,
and
how
to
structure
the
work
that
way
or
whether
you
want
to
like
have
a
bunch
of
PRS.
That's
all
merged
together
in
like
a
mega
branch
and
they
consequently
having
a
mega
branch
PR
that
other
PRS
are
descendants
from.
A
B
Probably
more
inclined
to
start
with
just
a
series
of
smaller
PRS:
that's
tends
to
work
better
with
hey
my
workflow,
but
be
also
with
my
kind
time
constraints.
I
probably
will
not
be
able
to
just
package
these
all
into
one
PR
in
a
reasonable
sort
of
timeframe.
I
feel
like
some
of
these
might
be
quick
hits
that
could
be
done
in
a
few
PRS,
but
particularly
writing
the
article
on
qo,
QoS
and
eviction
that
will
take
a
significant
amount.
F
So
one
thing
I
would
like
to
add
this
near
and
dear
to
my
heart.
As
of
the
past,
24-48
hours
is
as
we're
talking
about
moving
content
around
I.
Think
part
of
the
strategy
needs
to
be.
What
are
we
doing
with
the
old
placeholder
for
the
content,
whether
that's
a
redirect,
whether
it's
a
you
know?
How
do
we
clarify
that
this
information
is
moved,
I,
know
things
like
scheduling,
paints
and
toleration,
czar
hot
topics,
you
know,
I
know
the
research
for
the
CK
exams,
PKD
exam.
F
You
know,
after
breaking
the
site
last
night
I'm
a
little
bit
more
conscious
about
what
is
the
historical
state
of
it,
and
how
do
we
convey
that
information
and
I
think
this
is
a
great
plan?
I
love,
the
idea,
I
think
scheduling
is
kind
of
the
spaghetti,
monster
and
and
I
think
that
I
commend
the
efforts.
It's
a
strong
plus
one.
E
Yeah
Adam,
mrs.
Brown
I,
think
it's
a
great
idea.
I
feel
free
to
use
me
as
your
your
your
go
to
contact
for
this
and
I'll,
make
sure
I'm,
reviewing
and
and
the
way
is
reviewing
so
you're
getting
timely
reviews
and
and
I
like
your
approach
of
breaking
it
up
into
little
chunks.
That
just
makes
it
easier
on
everybody
doing
the
low
fruit
first
grab.
G
A
The
downside
to
using
an
alias
here
is
that
it
it
lays
it
preserves
the
existing
content
organization,
but
overlays
an
almost
invisible
mesh
of
reorganization
on
top
of
it
and
that
the
the
more
sensible
organization
isn't
the
one
that's
the
most
visible.
So
I
like
aliases
are
beautiful
things,
but
in
this
case
I
think
I.
A
A
G
Gonna
say
that
if
you're
gonna
do
even
if
you're
gonna
do
it
at
the
redirects
file
level,
what
I
would
maybe
do
Adam
is
like
go
through
the
actual
content,
changing
and
reorganization
in
one
pass,
and
once
you're
done
with
that,
I
would
then
open
another
PR
and
actually
like
manually
change.
The
links
like
do
us
do
a
search
and
replace,
and
then
second
PR,
once
all
the
content
work
is
done,
so
that
we
aren't
actually
relying
on
the
redirect
at
least
internally,
because
they
get
messy.
B
A
A
No,
that's
that's
very
smart.
Okay,
no
I
understand
that
yeah,
that's
yeah!
Let
let's
make
sure
that
the
sub
redirects
file
isn't
a
load-bearing
structure
when
first
rip
bro
can
link
specifically
yeah
good
catch
list.
Very
nice,
okay
Adam!
If
you
are
done
sharing,
do
you
mind
ending
your
share
I'll
gladly
share.
Thank
you
very
much
cheers
thanks.
That
was
really
cool,
I
think.
A
A
F
F
A
Talk
about
architecture,
so
we've
got
an
open
poll
requests,
18,
8,
1,
7
and
changes
the
build
scripts
for
how
we
actually
build.
What
the
the
build
commands
that
we
sense
and
that
liffe
I
to
actually
build
the
site-
and
this
is
a
little
bit
of
cleanup
work.
There
was
there's
another.
There
was
a
previous
PR,
it's
linked
to
from
1881
7
that
removed
the
been
it
know.
A
Basically,
we
no
longer
run
the
script
that
checks
for
hugo
version
mismatches
between
make
file
in
that
fly
that
tamil,
so
you
can
meet
when
seven
is
from
Tim
Bannister.
It's
basically
it's
a
straightforward,
clean
up
file,
but
Tim
wanted
to
discuss
it
in
the
meeting,
because
it
is
an
architectural
change
and
he
wanted
to
discuss
it
out
of
an
abundance
of
caution.
So
it
looks
good
to
me.
I
think.
A
F
So,
first
of
all,
I
don't
see
a
reason
why
it
shouldn't
merge,
but
I
have
ran
across
issues
where
I've
updated
notified
at
Tamil
to
have
a
different
network
aversion
or
the
different
hugo
version.
Where
then
in
the
build
would
fail,
and
it
would
kick
back
and
say,
here's
what
happened
say:
oh
shoot!
That's
right!
We've
got
the
two
locations
we
need
to
update.
I,
don't
have
the
historical
context
why
those
both
exist
if
I
had
to
make
an
educated
guess.
F
I
believe
that
this
goes
back
to
its
kind
of
is
related
to
a
question
earlier
to
local
testing.
I
believe
the
local
testing
advise
is
to
use
a
make
file
to
build
your
local
environment
and
I'll
have
to
double
check
that
so
I
believe
it's
a
matter
of
staying
in
sync
with
how
you're
doing
things
locally
versus
how
we're
doing
things
in
net
liffe
I
I.
F
Don't
think
this
will
have
an
impact
to
the
bills
per
se,
but
if
we
want
to
continue
down
with
that
recommendation
of
using
the
make
file
for
building
I'm,
not
sure
if
this
makes
sense,
but
I
believe
now
that
I'm
talking
out
loud.
This
change,
if
I'm
not
mistaken,
let
me
take
a
look
real
quick.
No,
this
isn't
the
one
there
was
another
proposed
change,
I
believe
that
basically
synchronize
the
two
files.
So,
instead
of
checking
it
just
went
off
of
one
source
of
truth,
and
this
looks
like
it
completely
removes
it.
A
That
imports
this
in
18
8
1
7,
there's
a
link
in
the
description
18020
and
that's
what
18020
does
is
it's
imports
in
to
make
file
the
Hyuga
version
as
a
variable
from
that
apply
that
tunnel?
So
that's
that's.
How
they're
kept
in
sync
now
is
because
I
think
we
both
did
that
at
least
two
times
each
was
updating,
Hugo
Hugo
version
in
one
file,
but
not
in
the
other,
and
what
18020
did
was
make
Matloff
I
taught
Tamil
the
source
of
truth
and
import
that
Hugo
Hugo
subversion
as
a
variable
into
the
make
file.
A
A
A
Architectural
same
thing:
okay,
a
quick
question
for
folks:
there's
a
PR
open
in
kubernetes
communities
right
now,
eighty
seven,
eight,
seven,
nine
and
what
this
PR
does
is
opens
it
basically
moves
changelog,
review
and
approval
out
of
the
root
directory
of
kaykai
and
puts
it
into
the
changelog
folder.
So
it
widens
the
pool
of
changelog,
reviewers
and
approvers,
and
this
is
basically
what
how
the
release
notes
happen
for
kubernetes
release
and
we
can,
if
we
like,
have
an
entry
block
in
the
owners
file
either
as
reviewers
or
approvers.
A
E
A
A
Let's
not
I
make
that
so
the
good
news
is
that
we
don't
have
to
take
any
action.
Cool,
okay,
diagrams
I,
see
no
David
Koepp
arose
again
today,
so
we
got
some
Jarrod
dropped,
a
link
and
feedback
to
the
diagram
statement
of
work
that
SEO
has
released
a
set
of
diagram
creation
guidelines
that
look
really
cool.
So
I
know
that
David
remark
that
he'd
seen
them
and
that
he
was
investigating
them
more
deeply
so
further
for
diagram
the
diagram
piece
to
move
forward.